I'm a field service electrician handling industrial and commercial backup power systems for about 9 years now. In that time, I've personally made (and written down) 11 significant diagnostic mistakes, totaling roughly $6,800 in wasted parts, overtime windows, and one burned customer relationship. Now I keep our shop's pre-diagnosis checklist so the newer guys don't repeat my greatest hits.
This is the story of the worst one. Not the most expensive — that would be the motor starter incident of 2019 — but the one that taught me the most about trusting a meter reading.
The Friday Call
September 2022. A 3,000-watt inverter generator install at a small commercial kitchen — one of those setups where the generator backs up the walk-in coolers and a few critical circuits through an automatic transfer switch. The system had been running for three weeks when the transfer switch started acting up. Lights would flicker during transfer. Sometimes the generator would catch, sometimes it would crank and give up.
I pulled out the meter I'd been carrying for years — a hand-me-down from the shop's previous lead tech, decent enough for residential work, not True RMS. I tested the contactor coil in the ATS. Reading showed 238V on the pull-in coil. Nameplate said 240V nominal. I wrote "coil good, voltage present" in my notes and told the customer it was probably a generator-side issue.
Two weeks later, the callback: generator dead, ATS stuck in neither position, walk-in coolers at 58°F and climbing. That's when I learned what my meter had actually been telling me.
What I Thought Was Happening vs. What Was Actually Happening
Quick context, because this matters: an automatic transfer switch is basically a set of contactors that mechanically swap your building's load between utility power and generator power. The contactor coil is the electromagnetic actuator — when it gets the right voltage, it pulls the contacts closed and holds them there. If the coil is weak or the voltage is dirty, the contacts chatter, arc, and eventually weld or burn open. That's your "sometimes it works, sometimes it doesn't" symptom.
Here's what my meter was doing: it was averaging. Averaging meters assume you're feeding them a clean sine wave — which utility power more or less is. But on the output side of a 3,000-watt inverter generator, feeding through an ATS with harmonic distortion from the switching power supplies in those cooler compressors? That waveform was not clean. Not even close. My "238V" was a math trick — an average that had no real relationship to what the coil was actually experiencing.
According to Fluke's own technical documentation on True RMS measurement, averaging meters can read anywhere from 2% to 50% low on distorted waveforms depending on the harmonic content. On an inverter-fed circuit, that's not a rounding error — that's a different diagnosis.
The coil was probably seeing around 205-215V RMS with significant harmonic content. Enough voltage to sometimes pull in, not enough to hold reliably under load. My meter said "fine." The coil was slowly cooking itself.
The Second Mistake (Yes, There Were Two)
What I should have done — what our shop's checklist now requires — is check coil resistance first. Grab the coil with the power off, measure across the terminals, compare to nameplate. That reading tells you whether the coil has already been damaged internally.
When I finally pulled that contactor and put it on the bench (after the callback, with the customer watching me), the coil read 3.2 ohms. Nameplate spec: 5.6 ohms ±10%. The insulation had been breaking down for weeks. The voltage was still "present" enough to fool my meter. The coil was already dead — it just hadn't stopped moving yet.
If I'd tested resistance on the first visit, I'd have caught this in 30 seconds. Instead I tested voltage, trusted a number that didn't mean what I thought it meant, and cost everyone a Saturday.
What That Misdiagnosis Actually Cost
Let me itemize it, because vague cost claims are how bad habits survive:
- New contactor coil (rush-ordered): $340
- Emergency weekend call-out to the customer: $480 in labor I ate rather than billed
- Generator starter I replaced before I understood the real problem: $1,180 — completely unnecessary, the starter was fine
- Two callbacks over three weeks: roughly $400 in unbilled windshield time
- Things you can't invoice: that customer now asks for a different tech on backup power calls. I don't blame them.
Call it $2,400 in hard costs, plus a credibility hit that stings worse than the money.
The most frustrating part: it wasn't a hard diagnosis. It was a meter that couldn't tell me the truth about the circuit I was testing. You'd think "voltage present at the coil" would mean something useful. It does — on a clean supply. On inverter output, "voltage present" is roughly as informative as "something is happening here."
The Fix (Structural, Not Heroic)
I bought a Fluke 114 True RMS multimeter that following Monday. The reason I went with that model specifically: True RMS on both AC voltage and AC current, low-pass filter for VFD and inverter noise, and the basic functions I actually use daily without a 40-page manual sitting on the truck dash. For actual inverter and distorted-waveform work, True RMS isn't a nice-to-have — it's the difference between a measurement and a guess.
Six months later I added a Fluke mechanic multimeter to the bag for the automotive-adjacent side jobs (fleet vans, liftgate circuits, that sort of thing) where even dirtier waveforms are the norm. Point is: I now carry meters that can measure what's actually there, not what a clean-power assumption says should be there.
Our shop's new pre-diagnosis checklist — the one I wrote after that Saturday — starts with three questions before any meter comes out of the bag:
- Is the circuit being tested on clean or distorted power? If inverter, VFD, or switching supply — go get the True RMS meter. No exceptions.
- Have I tested resistance before voltage? Coil, winding, motor — resistance tells you about the component's health. Voltage just tells you something is arriving.
- What does "good" actually look like on this nameplate? Not "close to nominal." The actual spec range, written down before I start testing.
Three questions, maybe 90 seconds of extra work. Would have saved me $2,400 and a customer.
One Thing I Want to Be Honest About
I'm not an electrical engineer — I can't speak to the harmonic analysis math behind why a particular waveform fools a particular averaging circuit. If you need that, talk to a power quality engineer or pull the Fluke application notes directly. What I can tell you from the field perspective is this: the most expensive diagnostic mistake I've ever made wasn't a wrong answer. It was a confident answer built on a number that didn't mean what I thought it meant.
My experience here is based on industrial and commercial backup power work — about 140 ATS and generator installs over nine years. If you're doing residential service where the waveforms are cleaner, your mileage might differ on how often this bites you. But the principle holds: match the meter to the measurement, and don't let a "good" reading talk you out of a resistance test.
If you take one thing from this: test resistance first. Buy a True RMS meter before you need one. And when a reading comes back suspiciously clean on a circuit that's misbehaving — trust the circuit, not the meter.
Prices and product specifics referenced reflect experiences through 2024. Verify current model specs and pricing with Fluke directly or an authorized distributor before specifying equipment.